Re: How Robots Will Steal Your Job

[Programmer Dude]

not yet. They can tell a dime from a nickel across a wide pool. We
can't come anywhere near that close.

Not yet, which is why I said "in principle". It's primarily an
engineering problem, since we know it *can* be done.

 

[Roedy Green]

here was a similar piece published in Nature:
http://www.nature.com/nsu/030818/030818-9.html

The problem was more trivial, but I gather it is a similar way
of sifting through and picking the best of an number of
possible possibilities.

 

[Constantinople]

here was a similar piece published in Nature:
http://www.nature.com/nsu/030818/030818-9.html

The problem was more trivial, but I gather it is a similar way
of sifting through and picking the best of an number of
possible possibilities.

But will it run Doom 3 and Half Life 2. That's what I really care about.

 

[Programmer Dude]

Do you mean the Turing Test?

Oops, yes.

I think that's not true. Eliza-like programs have passed the
Turing Test many times.

It depends a lot on the intelligence and knowledge of the test
person.

Sure. I just spent a facinating hour talking to A.L.I.C.E. I
don't know how she ranks, but she's obviously software.

 

[Wojtek]

And the pain and exhaustion they undergo is more than any CONSCIOUS
being would tolerate. They are non-aware.

Yet humans will submit to short term pain for long term gain, whereas
animals will not.

Example:
My cat hates wearing a collar. Yet the only way she is let outside is
with a collar. This is a long term relationship: collar = outside.

If she wants to go out to play, she sits by the door and waits until
someone notices. She does not get the collar, she does not sit by the
collar. In fact, she will sort of resist the collar being put on. Once
it IS on, she goes straight to the door and waits for it to be open
(proves that she knows about the relationship).


Humans undergo surgery to augment some part of their bodies. They KNOW
that post surgery pain will be present, yet they submit to it.

The Dune series even had an "animal test" where a human will make
themselves tolerate pain, knowing that it IS a test and they must pass
it, whereas an animal will immediately remove their hand from the
pain.

It made sense to me when I first read it, and it still makes sense to
(after many years of casual observation of animals).

BTW, children also resist short term pain, not being mature enough to
accept the pain for long term gain. This includes (IMHO) young adults
that do not study for tests but would rather play, not seeing the long
term benefits.

 

[Wojtek]

Wojtek,

creativity may consist of two components.

One is the ability to have new ideas. The computer equivalent is
the random number generator.

How does a mathematical formula that requires an external seed equate
to creativity? Since an identical seed will always have the same
"random" number set.

The second is the ability to sort the chaff from the wheat, to
filter out those few ideas that are genuinely useful. This
requires some intelligence.

Hmmm, by this definition a chess program has intelligence, since it
creates possible action paths, assigns a value to the outcome of each
path, then acts on the "best" action path.

As far as I know, Big Blue was not successful because it had much
better algorithms, but that it could "explore" more paths within time
constraints. Huge memory and massive parallel processing.

 

[Programmer Dude]

It is quite common for human languages to have words for 1 2 3 many.

Quite common? I've heard of it in some very primitive cultures that
haven't discovered math, yet, but I don't think it's common. It is
interesting, though, that we still have distinct words for 1, 2 and
3 (one, couple, few).

 

[Programmer Dude]

What about vet visits? Do some animals have a calendar good enough
to predict them?

No data. IME, our vet visits aren't regular enough. She enjoys
the vet (unlike some dogs), and she gets very excited when we get
about a block away.

One thing that always impressed me was that, when we'd go visiting,
she'd wake up and become alert and restless when we were only in
the general neighborhood of home (still several miles away). I
figure there must be a distinct smell to a region. (I've read that
the dog sense of smell is to them what our sense of sight is to us:
a primary sense. I frequently see her identify an old scrap of
food outside from more than 10 feet away.)

 

[Programmer Dude]

creativity may consist of two components.

One is the ability to have new ideas. The computer equivalent is
the random number generator.

The second is the ability to sort the chaff from the wheat, to
filter out those few ideas that are genuinely useful. This
requires some intelligence.

I would argue that even the first is usually intelligently
directed towards some goal and not entirely random.

 

[Programmer Dude]

I saw an extended demonstration.

Interesting. Got a cite I can check out?

 

[FM]

Interesting, that "selection" by itself implies some higher
authority which defines what is good and bad (criterion for selection,
someone or something which desides say short or long fins of carp to kill off).
By saying "natural" you did not remove the need of such authority,
you just made it undefined but still implied. So, what physical law
do you think defines direction of evolution?

Well, all of them. More specifically, the reason why
there seems to be a direction (towards greater complexity,
greater size, greater intelligenc, etc) is that for each
of those variables, there's a strict lower bound, zero,
and we started much closer to there than whatever upper
bound there might be. This isn't my idea - I read it
somewhere else - but it's pretty obvious from a
mathematical standpoint.

My answer is - it is the old good second law of thermodynamics,
which in fact defines existance of any process whatsoever.

Physical laws are generally descriptive, not prescriptive.
The second law of thermodynamics, in particular, isn't either.
It merely defines the notions of directionality of time and
entropy in relation to each other.

The requirement of second law of thermodynamics to Life
is "catalyze (accelerate) increase of entropy as much as you can".

There is no such requirement whatsoever. Life doesn't have to
do anything. Survival isn't a matter of increasing entropy.
And even if it were, it has nothing whatsoever to do with the
second law of thermodynamics.

In fact, you're saying over time, there will be more certainty
in exactly what kinds of organisms will survive, which appears
incorrect given the current level of biodiversity and implies
reduction in entropy (not strictly, but to be pedantic, nothing
you've ever said has anything to do with entropy).

Mechanism how this requirement is enforced is simple:
1) The catalysts which can not do this as fast as fundamentaly possible
are killed off - by advance of others which can do it faster and
pushing them from particular thermodynamical niche.

2) At the other hand, catalysts which are not able to use sufficent
part of releazed energy for own repair and propagation (regardless to
question how efficient they are as accelerators) just decompose
by themselfs, because any complex catalyst is not thermodynamically
stable.

In short, organisms that can use external resources to
reproduce rapidly will be more successful than those that
aren't as good at doing so. That has nothing to do with
thermodynamics and doesn't explain anything. Neither
complexity, nor faster metabolism contributes directly to
reproductive success.

This simple approach explains why Life increases its complexity
and even in what direction it evolves - in direction of higher energy
conversion (faster entropy increase) and higher ability to self-repair
and multiply.

Any change, be it on genetic level of single organism,
or on macro level (say, certain polytical system) which goes contrary to
this principle, is doomed to cause extintion.

There seem to be plenty of bacteria species around that
are nowhere near as complicated as various extinct species.
And that's not to mention viruses.

Dan.

 

[Roedy Green]

void magicTimeTravelUserInput()
{
do {
data_entry_form();
} while (!test_error());
}

That is nothing like it. See http://mindprod.com/jgloss/jaunting.html

That would not let you even back up a field in your data entry.

 

[Roedy Green]

Then it strikes me that these putative OOBEs are indistinguishable
from mere dreams.

The difference is the "feel". They feel like waking consciousness.
They have much more continuity than a dream. You remember just as
easily as waking consciousness, unlike dreams.

For the time being, I consider them a special form of dream, or
perhaps closer, hypnotic state.


For some people, they are quite dreamlike with visits to other
planets, or meetings with extraordinary beings. For others they are
quite mundane, just wandering around the house or office.

But then most of my dreams are about debugging Java code.

 

[Roedy Green]

poultry is smart too !!!

but not guinea fowl. Those things are too stupid to stay off a
highway.

 

[Roedy Green]

If a being has a model of the
world, at least of its surroundings, and if in this model the
being itself is included, then we call this conscience (or
self-awareness, which, to me, is about the same thing

Everything will a nervous system would fit that description would it
not? The nervous system models state both within and without the
animal. If an animal is smart enough not to chew off its own body
parts it must have some notion of self and other.

Dolphins presumably think in sounds. The have a great whacking lobes
in the brains for processing it.

A dolphin posing this question might define as intelligent a being
that kept an internal sonic model of its universe. Visual ones are
clearly inferior. You can't even tell the emotional state or health
of another purely visually. vision is an impoverished superficial way
of modeling reality.

 

[Roedy Green]

How does a mathematical formula that requires an external seed equate
to creativity? Since an identical seed will always have the same
"random" number set.

One theoretical answer to this is. Let's say your idea can be
summarised in a 4K paragraph. There are (2^16)^4096 possible
paragraphs. Each can be represented by an integer.

Intelligence is fishing around in that set to come up with the best
paragraphs.

One very brute force would be to generate a random number, evaluate
the idea. If it is "good" keep it, if not throw it back.

If you had some sort of quantum computer that could generate and
evaluate ideas in a reasonable length of time, surely that has to
count as intelligence.

 

[Roedy Green]

Appeared being a key word there.

We evolved from just such creatures.

Our motivations for food etc. could very well be identical. If it
works, why change it?

On the other hand, robots evolved via a totally different path and
thus there is no reason for them to share characteristics with us.


When you postulate different mechanisms for animals and humans what
you are really expressing is your species chauvinism. You don't give
any REASONS they should be different. You just want them different so
you can feel superior to animals and justify exploiting and
mistreating other species.

From a biological point of view we are very similar to chimps. We
stopped interbreeding with them not all that long ago on the
evolutionary tree. We are like nouveau riche putting on airs.

I suspect your religious training taught that humans are not animals,
and they it is highly shameful your were born naked, in bed, with a
woman.

 

[Roedy Green]

There is a book called A Whack On The Side of the Head. It comes with
a deck of cards used like a random number generator to make you look
at your problem from new angles.

Much of creativity comes from playfulness and curiosity. It comes from
combining two unlikely components.

With a new brainstorming idea, you don't stomp on it proving why it is
stupid. You look for any elements of it that can be salvaged for a
more practical solution.

 

[Wojtek]

One theoretical answer to this is. Let's say your idea can be
summarised in a 4K paragraph. There are (2^16)^4096 possible
paragraphs. Each can be represented by an integer.

Intelligence is fishing around in that set to come up with the best
paragraphs.

I think the key here is "best paragraphs". The fishing around part is
trivial.

One very brute force would be to generate a random number, evaluate
the idea. If it is "good" keep it, if not throw it back.

Yes, but randomly jumping around is not the intelligence part. The
evaluation is.

If you had some sort of quantum computer that could generate and
evaluate ideas in a reasonable length of time, surely that has to
count as intelligence.

What is considered reasonable time? Does it matter that it took us
(humans) many thousands of years to create the concept of modern math
(ie: the number zero as a place holder). Does that make us not
intelligent? Or are we intelligent because we did manage to do it at
all?


I think the crux of the matter is the ability to do things that are
not hard wired into us.

So yes, a computer program may "appear" to be intelligent because it
can come up with an answer to some question. BUT, if the program does
not have the algorithm already in place, it cannot cope with the
problem.

It took much effort to get robots to walk. And even now, a new type of
obstacle can flumox the robot. Yet any insect will be able to traverse
that new obstacle.

If a program can, *on its own*:
- come up with solutions to problems that have not been predetermined
by the programmer
- select solutions based not on exhaustive brute force inspection of
every possible solution, but on a "Eureka!" type of selection

then, I think, we can start attributing intelligence to the program
(or at least to the programmer).

 

[Roedy Green]

Yes, but randomly jumping around is not the intelligence part. The
evaluation is.

You need both. I think I would recognise a solution to global
problems if I saw one. However, I can't for the life of me think it
up. I don't have the compute power to generate enough alternatives
and evaluate them.

In some problems, the evaluation is the easy part: e.g. does this
transmission line design violate any of the safety design criteria?
How expensive would it be to build.

The hard part is fishing around in the sea of all possibilities
looking at the most promising possibilities since you don't have time
to consider them all. One technique for doing this is called dynamic
programming. I used it in Optow. We also used it managing water
levels in interconnected dams.